Delft University of Technology

Template-assisted bottom-up growth of nanocrystalline diamond micropillar arrays

Frota Sartori, Andre; Overes, Bart; Fanzio, Paola; Tsigkourakos, Menelaos; Sasso, Luigi; Buijnsters,Josephus G.DOI10.1016/j.diamond.2019.03.017Publication date2019Document VersionFinal published versionPublished inDiamond and Related Materials

Citation (APA)Frota Sartori, A., Overes, B., Fanzio, P., Tsigkourakos, M., Sasso, L., & Buijnsters, J. G. (2019). Template-assisted bottom-up growth of nanocrystalline diamond micropillar arrays. Diamond and Related Materials,95, 20-27. https://doi.org/10.1016/j.diamond.2019.03.017

Important noteTo cite this publication, please use the final published version (if applicable).Please check the document version above.

CopyrightOther than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consentof the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.

Takedown policyPlease contact us and provide details if you believe this document breaches copyrights.We will remove access to the work immediately and investigate your claim.

This work is downloaded from Delft University of Technology.For technical reasons the number of authors shown on this cover page is limited to a maximum of 10.

Contents lists available at ScienceDirect

Diamond & Related Materials

journal homepage: www.elsevier.com/locate/diamond

Template-assisted bottom-up growth of nanocrystalline diamond micropillararrays☆

André F. Sartori⁎, Bart H.L. Overes, Paola Fanzio, Menelaos Tsigkourakos1, Luigi Sasso,Josephus G. Buijnsters⁎

Department of Precision and Microsystems Engineering, Research Group of Micro and Nano Engineering, Delft University of Technology, Mekelweg 2, 2628 CD Delft, TheNetherlands

A R T I C L E I N F O

Keywords:Bottom-up growthMicropillar arraysCVD diamondImprint lithographyTemplate

A B S T R A C T

Micro-patterned diamond has been investigated for numerous applications, such as biomimetic surfaces, elec-trodes for cell stimulation and energy storage, photonic structures, imprint lithography, and others. Controlledpatterning of diamond substrates and moulds typically requires lithography-based top-down processing, which iscostly and complex. In this work, we introduce an alternative, cleanroom-free approach consisting of the bottom-up growth of nanocrystalline diamond (NCD) micropillar arrays by chemical vapour deposition (CVD) using acommercial porous Si membrane as a template. Conformal pillars of ~4.7 μm in height and ~2.2 μm in widthwere achieved after a maximum growth time of 9 h by hot-filament CVD (2% CH4 in H2, 725 °C at 10mbar). Inorder to demonstrate one of many possible applications, micropillar arrays grown for 6 h, with ~2 μm in height,were evaluated as moulds for imprint lithography by replication onto hard cyclic olefin copolymer (COC) andonto soft polydimethylsiloxane (PDMS) elastomer. The results showed preserved mechanical integrity of thediamond moulds after replication, as well as full pattern transfer onto the two polymers, with matching di-mensions between the grown pillars and the replicated holes. Prior surface treatment of the diamond mould wasnot required for releasing the PDMS replica, whereas the functionalisation of the diamond surface with a per-fluorododecyltrichlorosilane (FDDTS) anti-stiction layer was necessary for the successful release of the COCreplica from the mould. In summary, this paper presents an alternative and facile route for the fabrication ofdiamond micropillar arrays and functional micro-textured surfaces.

1. Introduction

Diamond is a material with desirable mechanical and chemicalproperties for biomimetic structures [1–3], (bio)electrochemistry [4],imprint lithography [5] and other applications, due to its high Young'smodulus (up to ~106 Nmm−2), high thermal conductivity(> 2000Wm−1 K−1) [4], low thermal expansion coefficient (10−6 K−1

at 300 K) [6], low coefficient of friction [7], non-toxicity, and versatilesurface chemistry. Many of those applications require patterning of thediamond surface into regular functional structures, such as nanowires[8,9], micro/nanopillars, microchannels [10] and pads [11], which arenormally achieved by multistep lithography (photo-, e-beam, focusedion beam) and/or etching processes [12,13]. For this reason, new fab-rication routes aimed at reducing complexity and cost have been in-creasingly investigated. These include, for example, the formation of

periodic surface structures by laser for the creation of photonic struc-tures [14] and enhanced electrodes [15], the combination of diamondwith foreign scaffold materials to enhance the surface area in electrode[16], biosensing [17] and field-emission [18] applications, the selectiveseeding of diamond nanoparticles through micropipetting [19], inkjetprinting [20,21] and surface functionalisation [22], and the use ofporous alumina templates [23,24] for the bottom-up growth of dia-mond micro/nanostructures.

The latter approach is attractive, since it shifts the need for pat-terning to materials which are easier to process and which enjoy a moremature and well-established industry. An obvious material candidatefor that purpose is anodised aluminium oxide (AAO), which has beenpreviously applied as etching masks for the fabrication of porous dia-mond membranes [23], as well as a template for growth of diamondnanopillars [24]. However, the pore size of AAO is limited to the sub-

https://doi.org/10.1016/j.diamond.2019.03.017Received 8 February 2019; Received in revised form 25 March 2019; Accepted 25 March 2019

☆ The authors declare no competing financial interest.⁎ Corresponding authors.E-mail addresses: A.FrotaSartori@tudelft.nl (A.F. Sartori), J.G.Buijnsters@tudelft.nl (J.G. Buijnsters).

1 Present address: Department of Applied Physics, National Technical University of Athens, Iroon Polytechniou 9, 15780 Athens, Greece.

Diamond & Related Materials 95 (2019) 20–27

Available online 26 March 20190925-9635/ © 2019 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/BY-NC-ND/4.0/).

T

micrometre range, the pore distribution lacks long-range order, and themembranes can be an undesired source of contaminants, such as Al andchemical residuals from the anodising process, into the chemical va-pour deposition (CVD) chamber. Silicon, on the other hand, can bemicropatterned on an industrial scale and does not have the drawbacksof AAO. For those reasons, we devoted this work to the study of bottom-up growth of diamond micropillars utilising commercially available,pre-patterned silicon templates. The resulting diamond microstructureswere employed as moulds for imprint lithography to demonstrate one ofmany possible applications.

Micro-/nanoimprint lithography is a high-throughput, high-resolu-tion, and low-cost contact patterning method used to create micro-fluidic devices and optical components, and to overcome resolutionlimits with photolithography [25]. The surface pattern of a mould isreplicated onto a thermoplastic material or onto a resist by mechanicalcontact and three-dimensional material displacement. The moulds canbe exposed to elevated pressures and temperatures and need to be usedrepeatedly and reproducibly. Diamond is an attractive material forimprint lithography moulds, due to its high hardness, high thermalconductivity, low thermal expansion coefficient, chemical inertness andwear resistance. Up until now, diamond stamps were fabricated usingsingle crystal or polished CVD diamond, which are micro-structured byfocused ion beam, reactive ion etching or e-beam lithography, wherescalability and cost are an issue [5,26–28]. Thus, template-basedbottom-up growth of diamond microstructures may enable the pro-duction of diamond moulds on wafer scale and help bring costs down.

In this work, we show the steps towards the successful fabrication ofnanocrystalline diamond micro-pillar arrays on silicon and their re-plication onto two different polymers: cyclic olefin copolymer (COC)developed by TOPAS®, and polydimethylsiloxane (PDMS), which areimportant materials for the fabrication of microfluidic, lab-on-chip andbiosensing devices [29,30]. The mould replication onto COC was per-formed by hot-embossing at high pressure and temperature, while thereplication onto PDMS was carried out at ambient conditions. Thepattern transfer was analysed by a combination of 3D optical profilo-metry and scanning electron microscopy (SEM).

2. Experimental

2.1. Diamond micropillar growth

Fig. 1 illustrates the fabrication steps of the diamond micropillarmaster mould. In a first step, 1) a ~1.5× 1.5 cm2 piece of bare siliconwafer of 0.4mm thickness (with native oxide) was spin-seeded withdetonation nanodiamond particles, according to the procedure de-scribed elsewhere [31]. Then, 2) the seeded wafer was loaded into ahome-built hot-filament chemical vapour deposition (HFCVD) chamberfor the growth of a thin layer of nanocrystalline diamond (NCD) of~0.2 μm thickness for 15min, performed at our standard conditions:725 °C, 10mbar, with 6 sccm CH4 and 300 sccm H2. Subsequently, 3) aporous Si membrane (1×1 cm2, ~17 μm thick, with square~2.5×2.5 μm2 pores arranged in a hexagonal pattern with ~4.2 μmpitch) from SmartMembranes GmbH, Germany, was placed onto thegrown NCD layer, and supported at the corners with additional piecesof Si. 4) The Si/NCD/membrane wafer was then loaded into the HFCVDchamber for further growth for up to 9 h. After that, 5) the membranewas mechanically and chemically removed, in order to expose thenewly formed diamond micropillars. For short growth duration, beforemechanical interlocking between the pillars and the membrane holesoccurs, ultrasonication in acetone was sufficient to remove the mem-brane. For longer growth duration, when the pillars are interlockedwith the membrane, only parts of the membrane can be pulled offmechanically (e.g. with a tweezer). The chemical process was thus ne-cessary to etch away the remaining pieces of the membrane, and con-sisted of dipping the sample in a bath of concentrated, boiling KOH, fora few minutes, until the membrane had been removed completely. The

process was completed with cleaning in a boiling concentrated mixtureof HCl+H2SO4+HNO3, which removed remaining carbon depositsand promoted the oxygenation of the diamond surface as a side-effect.The samples were rinsed in deionised water as a final step.

2.2. Mould replication onto COC and PDMS

Fig. 2 illustrates the replica moulding procedures. Replication intopolydimethylsiloxane (PDMS) was achieved by casting a PDMS solution(10 parts of elastomer to 1 part curing agent, w/w ratio) onto thediamond master mould and letting it cure at 70–80 °C for at least 1 h.After that, the PDMS replica was peeled off from the diamond mouldand further analysed. Mould replication onto cyclic olefin copolymer(COC) was achieved by hot-embossing the diamond micropillar mastermould onto a 1mm thick COC wafer, grade 6013 from TOPAS®, in anEVG510 wafer bonder setup equipped for the purpose. The procedurewas performed in vacuum, with a load of ~2500 N applied for 8min, at170 °C (above the COC's glass transition temperature of 130 °C). Afterthis time, the sample was let cool down to room temperature, still underload, and released thereafter. The COC was manually demoulded. Inorder to facilitate the latter step and avoid breaking the thin siliconwafer, the diamond micropillar master mould was previously coatedwith a FDDTS anti-stiction layer, as described below.

2.3. 1H, 1H, 2H, 2H-Perfluorododecyltrichlorosilane (FDDTS) anti-stictionlayer

The diamond micropillar master moulds were coated with FDDTSbefore hot-embossing into COC. The acid cleaning treatment described

1) Spin seedingDiamond seeds

Si

2) Diamond growthDiamond film

Si

3) Membrane placement

Si

Porous membrane

4) Diamond growth

Si

Porous membrane

5) Membrane removalDiamond pillars

SiFig. 1. Fabrication steps of the diamond micropillar master mould.

A.F. Sartori, et al. Diamond & Related Materials 95 (2019) 20–27

21

earlier produced the necessary oxygenation of the diamond surface forthe binding of the FDDTS molecules. 1H, 1H, 2H, 2H-Perfluorododecyltrichlorosilane (97% purity) was obtained fromSigma-Aldrich, and diluted in ethanol (1% w/w ratio). The mouldswere dipped into the solution and ultrasonicated for 30 s, then leftresting for 30min. After that, the moulds were taken out and ultra-sonicated in ethanol for 30 s, followed by baking at ~130 °C on a hotplate for 30min.

2.4. Structure and morphology characterisation

Scanning electron microscopy (SEM) measurements were performedwith a JEOL JSM-6010LA scanning electron microscope, at 5 keV andsecondary electron detector, and with a field-emission FEI NovaNanoSEM 450 setup operating at 15 keV, with a high-resolution (im-mersion mode) secondary electron detector. Raman spectroscopymeasurements were performed with a Horiba LabRAM HR setup,equipped with an argon ion laser operating at 514 nm, and spectralresolution of ~0.3 cm−1. Height profile imaging of the patternedsamples was performed with a Bruker Contour GT-K 3D optical profil-ometer operating at highest (100×) magnification, with white lightillumination, and VXI mode for best accuracy.

3. Results and discussion

3.1. Template characterisation

Fig. 3 shows SEM micrographs of the porous silicon membrane usedas a template for the growth of diamond micropillars. Apart fromconfirming the dimensions specified by the supplier (see Section 2.1above), notable characteristics of the template were the bottleneck atabout ~1 μm depth into the pores, which is ~1.8 μm wide, whereas theoverall pore diameter reaches up to 2.5 μm. Fig. 3(c) shows how thesquare pores morph into a more rounded shape along the depth of themembrane.

3.2. Diamond micropillars

Several diamond micropillar samples were produced, with differentgrowth durations from 3 to 9 h. Before placing the Si template, a~0.2 μm thick NCD layer was grown in order to ensure a closed dia-mond film at the bottom. The template was then placed onto it, with theflat front side facing down, and the diamond growth was resumed. Afterremoving the template, the produced samples were imaged by SEM(Fig. 4(a–c)) and analysed by Raman spectroscopy (Fig. 4(d)). The SEMimages show that diamond grows selectively into the pores of thetemplate, leading to very well-defined, uniform and well-aligned mi-crostructures. As expected from the pore shape, the pillars evolved fromsquares to rounded squares, and are narrow near the base, matching thebottleneck inside the pores of the template shown in Fig. 3(b). Fig. 4(e)shows an optical microscopy image of the sample surface with NCDmicropillars grown for 9 h. The iridescent area at the centre corre-sponds to two growth zones (as indicated by the numbers): (1) fullygrown pillars and (2) pillar height gradient. Outside the iridescent area,in zone (3), the NCD layer grew flat. The reason for this variation isexplained in Fig. 5 below. The Raman spectrum of a random micropillarafter 6 h growth is compared in Fig. 4(d) with the spectrum measuredoutside the template area (i.e. flat NCD layer), as indicated by the insetSEM images. The combination of the well-known diamond one-phononline at 1333 cm−1 [32] with prominent D- and G-bands confirms that

PDMS

Casting + curing

Si

COC

Hot-embossing

Si

3) Demoulding

PDMS / COC

Si

2) Mould replication

Si

1) Surface treatment

FDDTS-coated diamond

Si

Oxygenated diamond

PDMS TOPAS® COC

Fig. 2. Fabrication steps of the PDMS and COC mould replicas.

Fig. 3. SEM micrographs of the porous silicon membrane fromSmartMembranes GmbH used to synthesise the NCD micropillars. (a) Tiltedsurface. (b) Cross section. The flat front side faced the initial NCD layer for themicropillar growth. (c) Backside. The nominal dimensions are: membrane sizeof 1× 1 cm2, ~17 μm thickness, pore diameter of up to ~2.5 μm and centroiddistance (pitch) of ~4.2 μm.

A.F. Sartori, et al. Diamond & Related Materials 95 (2019) 20–27

22

the material on the pillar is diamond, with a high percentage of sp2

carbon [33–35]. The flat NCD film grown freely outside the membraneshows relatively lower sp2 carbon content, as seen from the less intenseG-band and virtually absent D-band. The observed difference is due tothe change in local growth environment around the template pores,which, on the one hand, partially inhibit the growth species (atomichydrogen and CxHy radicals) from reaching the diamond surface, thusdecreasing the local growth rate, and, on the other hand, likely serve asa local source of carbon from deposits on the membrane. The overalleffect on the composition is comparable to growing NCD with higher[CH4]/[H2] ratio. In addition to the overall shape, the Raman spectrum

on the pillar also shows the presence of peaks related to trans-poly-acetylene, which are common signatures in nanocrystalline diamond[35,36]. The silicon peak at 520.7 cm−1 arises from the underlying Sisubstrate. Overall, the SEM and Raman results show that the NCDmaterial grew homogeneously and conformally into the pores.

Fig. 5 illustrates the development of the NCD micropillars throughthe template pores as a function of time, but also as a function of thegap between the Si template and the NCD substrate. The left column(zero gap) corresponds to the fully patterned growth zone where themicropillars started to grow immediately from the beginning of theCVD growth process, which is the situation shown in Fig. 4(a–c),

Fig. 4. (a–c) SEM micrographs of the tem-plate-grown NCD micropillars after growthfor (a) 3 h, (b) 6 h and (c) 9 h. Note how thebottleneck inside the pores of the template(Fig. 3(b)) shaped the base of the diamondpillars after 6 h. (d) Raman spectra (as-measured) taken from the top of the mi-cropillar (green curve) and from outside themembrane/patterned area (black curve) forcomparison. (e) Optical microscopy imageof the 9 h sample, showing distinct growthzones indicated by the numbers (1–3). Thecomplete image corresponds to the1×1 cm2 area which was covered by the Sitemplate. The dark areas at the upper-leftand bottom-right corners match the loca-tion of the supporting Si pieces used to holdthe membrane in place during the growthprocess.

Fig. 5. Illustration of how the NCD micropillars developduring growth through the template pores as a function ofgrowth time (vertical direction) and the gap between the pre-grown flat NCD layer and the Si template (horizontal direc-tion). The zero-gap column on the left (shaded grey back-ground) corresponds to the samples shown in Fig. 4(a–c), witht1= 3 h, t2= 6 h and t3= 9 h. The t2 line shows how thepillar height decreases with increasing template gap on the6 h sample.

A.F. Sartori, et al. Diamond & Related Materials 95 (2019) 20–27

23

respectively for t1= 3 h, t2= 6 h and t3= 9 h. Fig. 5 also shows SEMimages of the 6 h sample as a function of template gap in the t2 row. Asthe template gap increases due to buckling, because the Si membrane isvery thin (~17 μm), the additional time it takes for the underlying NCDlayer to reach the base of the template before NCD pillars can developthrough the pores, leads to smaller pillar height. In the extreme casewhere the gap is too large for duration of the NCD growth process, thediamond remains flat in that zone. The resulting optical effect of suchvariations on the NCD micropillar formation is shown in Fig. 4(e) forthe 9 h sample, as described earlier.

3.3. COC replica

After growth for 6 h, the diamond micropillar master mould wascoated with a FDDTS anti-stiction layer, and then hot-embossed intoCOC (negative replica). After cooling down to near room temperature,the replica was mechanically detached from the diamond master mouldand used as a mould for replication onto PDMS (positive replica), inorder to reproduce the shape of the diamond micropillars from the COCreplica. The three samples (i.e., the diamond master mould and the twopolymer replicas) were analysed by 3D optical profilometry, as shownin Fig. 6(a–c). SEM images of the COC and PDMS replicas are shown inFig. 7(a,b), respectively. It can be seen in Fig. 6(a–c) that the diamondmicropillars resulted homogeneous in both lateral and vertical dimen-sions, whereas the COC replica resulted in a somewhat less regularpattern, with slight distortion of the hole perimeter. Extra irregularitieson the top surface can also be seen. On the other hand, the PDMS re-plica shows pillars matching the original diamond master mould veryclosely in both lateral and vertical dimensions. Vertical and horizontalline profiles of the three areas (taken along the indicated dashed lines)are compared in Fig. 6(d). They show a clear match between the pillarsand the holes, and a full imprint into the ~2 μm inter-pillar depth. A

statistical distribution obtained from particle analysis of the three top-view images is given in Fig. 6(e). The results are summarised in Table 1.The area corresponds to either the top surface of the pillars or thebottom surface of the holes.

The statistical analysis of the optical profilometry images shows amismatch between the area of the top of the pillars of the diamondmaster mould and that of the bottom of the holes in the COC negativereplica, with the holes in the COC ~24% smaller than the pillars on themould, and with ~2.2% smaller nearest neighbour distance.Considering that the COC cooled down from 170 °C (above its glasstransition temperature) to 25 °C, and taking into account its thermalexpansion coefficient, εCOC= 0.6×10−4 K−1, the expected contrac-tion purely due to thermal expansion is only 0.90%. Thus, although thismay have contributed to the observed 2.2% shrinkage in length,thermal shrinkage alone cannot explain the relatively smaller size of theholes. The statistical analysis shows, however, a very good match be-tween the PDMS positive replica and the diamond master mould, withmatching dimensions between pillars and holes, and retained averagenearest neighbour distance from the COC. This suggests that the area ofthe COC holes measured by optical profilometry was underestimated,mostly likely due to optical artefacts at the edges of the structures fromthe limited lateral resolution of the setup. This is confirmed by the SEMimage of the COC, shown in Fig. 7(a), where the hole dimensions matchthose of the original mould. Thus, overall, the replication of the dia-mond master mould onto COC was successful.

3.4. PDMS replica

After growth for 6 h, a diamond micropillar master mould was re-plicated onto PDMS as illustrated in Fig. 2, resulting in an imprintedstructure similar to the COC replica discussed above, as shown inFig. 8(a,b). Fig. 8(c) compares the line profiles of the two samples in

(e)

10 µm

10 µm

3 4 50

20

40

60

80

100

120

AREA (µm2)

Diamond mould (positive)COC replica (negative)PDMS replica (positive)

0 10 20 30 400

1

2

3

4

HEI

GH

T (µ

m)

LENGTH (µm)

0

1

2

3

4

HEI

GH

T (µ

m)

Diamond mould (positive) (d)

Vertical

Horizontal

(a) Diamond mould (positive)

(b) COC replica (negative)

0.0 3.0 µm1.0 2.0

10 µm

(c) PDMS replica (positive)

COU

NT

COC replica (negative)PDMS replica (positive)

Fig. 6. (a–c) 3D optical profilometry image of (a) the 6 hdiamond master mould, (b) the COC replica, and (c) the PDMSreplica of the COC replica. (d) Comparison of the vertical andhorizontal line profiles, taken from (a), (b) and (c) as in-dicated by the white dashed lines. (e) Histogram (n=33) ofthe area of the top of pillars from (a,c) and of the bottom ofthe holes from (b).

A.F. Sartori, et al. Diamond & Related Materials 95 (2019) 20–27

24

both vertical and horizontal directions, as indicated by the dashed linesin Fig. 8(a,b). Fig. 8(d) shows histograms of the areas of both the pillarsand the holes, derived from statistical analysis of the 3D optical profi-lometry images. The results are given in Table 2, and show full

replication of the mould onto the PDMS, with similar height/depth andlateral dimensions between the pillars and the holes. Nearest neighbouranalysis resulted in ~1% contraction of the PDMS replica, relative tothe diamond mould, which explains the minor offset seen between the

2 µm 2 µm

5 µm

(a) COC negative replica (6h mould)

5 µm

(b) PDMS positive replica (6h mould)

(c) PDMS negative replica (6h mould) (d) PDMS negative replica (9h mould)

Fig. 7. Tilted (30°) SEM micrographs of (a) the COC negativereplica shown in Fig. 6(b), (b) the PDMS positive replica (seeFig. 6(c)) of the COC replica shown in (a), (c,d) PDMS ne-gative replicas of the diamond master moulds grown for (c)6 h and (d) 9 h, respectively. The yellow outlines in (c) and(d) are a guide for the eye to highlight the difference in poredepth between the two samples. (For interpretation of thereferences to color in this figure legend, the reader is referredto the web version of this article.)

Table 1Summary of the statistical analysis of the 3D optical profilometry images shown in Fig. 6. The given numbers are average values ± standard deviation (σ).

Sample Area ± σ (μm2) Square length ± σ (μm) Nearest neighbour distance ± σ (μm)

Diamond master mould, 6 h 5.00 ± 0.14 2.24 ± 0.03 4.11 ± 0.02COC (negative) replica 3.81 ± 0.33 1.95 ± 0.03 4.02 ± 0.06PDMS (positive) replica 4.99 ± 0.22 2.23 ± 0.05 4.02 ± 0.04

AREA (µm2)0 10 20 30 40

0

1

2

3

4

HEI

GH

T (µ

m)

LENGTH (µm)

0

1

2

3

4

HEI

GH

T (µ

m)

Diamond mould(PDMS)

PDMS replica

3 4 50

20

40

60

80

100

120

CO

UN

T

Diamond mould(PDMS)

PDMS replica

10 µm 10 µm

(b) PDMS replica

(c)

(a) Diamond mould

Vertical

Horizontal

(d)

Fig. 8. (a,b) 3D optical profilometry image of (a)the 6 h diamond mould and of (b) its PDMS re-plica. (c) Comparison of the vertical and hor-izontal line profiles, taken from (a) and (b) asindicated by the white dashed lines. (d)Histogram (n=25) of the area of the top of pil-lars from (a) and of the bottom of the imprintedholes from (b).

A.F. Sartori, et al. Diamond & Related Materials 95 (2019) 20–27

25

line scans shown in Fig. 8(c). The SEM image of the PDMS replica inFig. 7(c) resembles that of the original Si template (Fig. 3(a)) with itsdistinct trapezoidal profile until the bottleneck, albeit smoothed. Theresults show that, like for COC, the diamond mould was fully replicatedinto the PDMS, with the resulting holes displaying matching depth tothe pillar height.

For comparison, the 9 h diamond micropillar master mould shownin Fig. 4(c) was also replicated onto PDMS, resulting in the structureshown in Fig. 7(d). It can be seen that the replica is very similar to thereplica of the 6 h mould (Fig. 7(c)), but with deeper holes, as high-lighted by the yellow outlines. The compliance of the PDMS made itsrelease from the 9 h mould possible, despite the bottleneck shape of themicropillars. Such approach was not feasible with COC, as a result ofthe higher stiffness of the copolymer material.

In summary, the results show that bottom-up fabrication of NCDmicropillars can be achieved with the use of commercial porous Simembranes as growth templates, with potential use as biomimetic de-vices, enhanced electrodes, or moulds for imprint lithography. Thelatter was demonstrated above, with the successful application of theNCD micropillars for imprint lithography onto COC and PDMS.

4. Conclusion

In this work, we demonstrated the bottom-up growth of nanocrys-talline diamond micropillars by chemical vapour deposition using amicroporous Si template mask. Conformal NCD pillars were grown in-side the ~2.2 μm square pores up to a height of ~4.7 μm for 9 h growth.The NCD micro-pillars grown for 6 h (~2 μm in height) were applied asmoulds for imprint lithography on COC (hard polymer) and PDMS (softpolymer), resulting in full mould replication in both cases. Replicationof taller pillars into COC was limited due to mechanical interlockingcaused by the bottleneck shape of the pillars, which was inherited fromthe commercial Si membrane. For smaller pillars, replication onto COCwas only possible with the addition of a fluorine-based anti-stictionlayer, whereas for replication onto PDMS no additional surface treat-ment was necessary. The presented template-assisted method for thegrowth of NCD microstructures may find important applications indiamond-based biomimetic and photonic devices, enhanced (bio)elec-trodes, as well as moulds for imprint lithography.

Acknowledgements

The authors thank the PME technical team for the lab support.

References

[1] R. Manai, E. Scorsone, L. Rousseau, F. Ghassemi, M. Possas Abreu, G. Lissorgues,N. Tremillon, H. Ginisty, J.C. Arnault, E. Tuccori, M. Bernabei, K. Cali, K.C. Persaud,P. Bergonzo, Grafting odorant binding proteins on diamond bio-MEMS, Biosens.Bioelectron. 60 (2014) 311–317, https://doi.org/10.1016/j.bios.2014.04.020.

[2] E. Granados, M.M. Calderon, J. Krzywinski, E. Wörner, A. Rodriguez,M.G. Aranzadi, S.M. Olaizola, Enhancement of surface area and wettability prop-erties of boron doped diamond by femtosecond laser-induced periodic surfacestructuring, Opt. Mater. Express. 7 (2017) 3389–3396, https://doi.org/10.1364/OME.7.003389.

[3] Z. Tian, W. Huang, B. Xu, X. Li, Y. Mei, Anisotropic rolling and controlled chiralityof nanocrystalline diamond nanomembranes toward biomimetic helical frame-works, Nano Lett. 18 (2018) 3688–3694, https://doi.org/10.1021/acs.nanolett.8b00828.

[4] S. Szunerits, C.E. Nebel, R.J. Hamers, Surface functionalization and biological ap-plications of CVD diamond, MRS Bull. 39 (2014) 517–524, https://doi.org/10.1557/mrs.2014.99.

[5] J. Taniguchi, Y. Tokano, I. Miyamoto, M. Komuro, H. Hiroshima, K. Kobayashi,T. Miyazaki, H. Ohyi, Preparation of diamond mold using electron beam litho-graphy for application to nanoimprint lithography, Jpn. J. Appl. Phys. 39 (2000)7070–7074, https://doi.org/10.1143/JJAP.39.7070.

[6] S. Stoupin, Y. V Shvyd'ko, Ultraprecise studies of the thermal expansion coefficientof diamond using backscattering x-ray diffraction, Phys. Rev. B. 83 (2011) 104102.

[7] J.G. Buijnsters, M. Tsigkourakos, T. Hantschel, F.O.V. Gomes, T. Nuytten, P. Favia,H. Bender, K. Arstila, J.-P. Celis, W. Vandervorst, Effect of boron doping on the wearbehavior of the growth and nucleation surfaces of micro- and nanocrystalline dia-mond films, ACS Appl. Mater. Interfaces 8 (2016) 26381–26391, https://doi.org/10.1021/acsami.6b08083.

[8] S. Szunerits, Y. Coffinier, E. Galopin, J. Brenner, R. Boukherroub, Preparation ofboron-doped diamond nanowires and their application for sensitive electrochemicaldetection of tryptophan, Electrochem. Commun. 12 (2010) 438–441, https://doi.org/10.1016/j.elecom.2010.01.014.

[9] P.W. May, M. Clegg, T.A. Silva, H. Zanin, O. Fatibello-Filho, V. Celorrio,D.J. Fermin, C.C. Welch, G. Hazell, L. Fisher, A. Nobbs, B. Su, Diamond-coated“black silicon” as a promising material for high-surface-area electrochemical elec-trodes and antibacterial surfaces, J. Mater. Chem. B 4 (2016) 5737–5746, https://doi.org/10.1039/C6TB01774F.

[10] A. Karczemska, Diamond materials for microfluidic devices, in: Diamond-basedMater, Woodhead Publishing, Biomed. Appl, 2013, pp. 256–271, https://doi.org/10.1533/9780857093516.2.256.

[11] R. Ramaneti, K.J. Sankaran, S. Korneychuk, C.J. Yeh, G. Degutis, K.C. Leou,J. Verbeeck, M.K. Van Bael, I.N. Lin, K. Haenen, Vertically aligned diamond-gra-phite hybrid nanorod arrays with superior field electron emission properties, APLMater. 5 (6) (2017) 066102, , https://doi.org/10.1063/1.4985107.

[12] P.A. Nistor, P.W. May, Diamond thin films: giving biomedical applications a newshine, J. R. Soc. Interface 14 (2017), https://doi.org/10.1098/rsif.2017.0382.

[13] S. Castelletto, L. Rosa, J. Blackledge, M.Z. Al Abri, A. Boretti, Advances in diamondnanofabrication for ultrasensitive devices, Microsystems Nanoeng. 3 (2017) 17061,https://doi.org/10.1038/micronano.2017.61.

[14] E. Granados, M. Martinez-Calderon, M. Gomez, A. Rodriguez, S.M. Olaizola,Photonic structures in diamond based on femtosecond UV laser induced periodicsurface structuring (LIPSS), Opt. Express 25 (2017) 15330–15335, https://doi.org/10.1364/OE.25.015330.

[15] A.F. Sartori, S. Orlando, A. Bellucci, D.M. Trucchi, S. Abrahami, T. Boehme,T. Hantschel, W. Vandervorst, J.G. Buijnsters, Laser-Induced Periodic SurfaceStructures (LIPSS) on heavily boron-doped diamond for electrode applications, ACSAppl. Mater. Interfaces 10 (2018) 43236–43251, https://doi.org/10.1021/acsami.8b15951.

[16] F. Gao, C.E. Nebel, Diamond-based supercapacitors: realization and properties, ACSAppl. Mater. Interfaces 8 (2016) 28244–28254, https://doi.org/10.1021/acsami.5b07027.

[17] G. Piret, C. Hébert, J.-P. Mazellier, L. Rousseau, E. Scorsone, M. Cottance,G. Lissorgues, M.O. Heuschkel, S. Picaud, P. Bergonzo, B. Yvert, 3D-nanostructuredboron-doped diamond for microelectrode array neural interfacing, Biomaterials. 53(2015) 173–183, https://doi.org/10.1016/j.biomaterials.2015.02.021.

[18] Y. Zou, P.W. May, S.M.C. Vieira, N.A. Fox, Field emission from diamond-coatedmultiwalled carbon nanotube “teepee” structures, J. Appl. Phys. 112 (2012)044903, , https://doi.org/10.1063/1.4748336.

[19] A.C. Taylor, R. Edgington, R.B. Jackman, Patterning of nanodiamond tracks andnanocrystalline diamond films using a micropipette for additive direct-write pro-cessing, ACS Appl. Mater. Interfaces 7 (2015) 6490–6495, https://doi.org/10.1021/am507900a.

[20] N.A. Fox, M.J. Youh, J.W. Steeds, W.N. Wang, Patterned diamond particle films, J.Appl. Phys. 87 (2000) 8187–8191, https://doi.org/10.1063/1.373516.

[21] A.F. Sartori, P. Belardinelli, R.J. Dolleman, P.G. Steeneken, M.K. Ghatkesar,J.G. Buijnsters, Inkjet-printed high- Q nanocrystalline diamond resonators, Small.15 (2019) 1803774, https://doi.org/10.1002/smll.201803774.

[22] P. Pobedinskas, G. Degutis, W. Dexters, W. Janssen, S.D. Janssens, B. Conings,B. Ruttens, J. D'Haen, H.-G. Boyen, A. Hardy, M.K. Van Bael, K. Haenen, Surfaceplasma pretreatment for enhanced diamond nucleation on AlN, Appl. Phys. Lett.102 (2013) 201609, https://doi.org/10.1063/1.4807591.

[23] H. Masuda, M. Watanabe, K. Yasui, D. Tryk, T. Rao, A. Fujishima, Fabrication of ananostructured diamond honeycomb film, Adv. Mater. 12 (2000) 444–447, https://doi.org/10.1002/(SICI)1521-4095(200003)12:6<444::AID-ADMA444>3.0.CO;2-K.

[24] H. Masuda, T. Yanagishita, K. Yasui, K. Nishio, I. Yagi, T.N. Rao, A. Fujishima,Synthesis of well-aligned diamond nanocylinders, Adv. Mater. 13 (2001) 247–249,https://doi.org/10.1002/1521-4095(200102)13:4<247::AID-ADMA247>3.0.CO;2-H.

[25] M.C. Traub, W. Longsine, V.N. Truskett, Advances in nanoimprint lithography,Annu. Rev. Chem. Biomol. Eng. 7 (2016) 583–604, https://doi.org/10.1146/annurev-chembioeng-080615-034635.

[26] J. Taniguchi, Y. Tokano, I. Miyamoto, M. Komuro, H. Hiroshima, Diamond

Table 2Summary of the statistical analysis of the 3D optical profilometry images shown in Fig. 8. The given numbers are average values ± standard deviation (σ).

Sample Area ± σ (μm2) Square length ± σ (μm) Nearest neighbour distance ± σ (μm)

Diamond mould, 6 h 4.26 ± 0.29 2.06 ± 0.07 4.08 ± 0.04PDMS (negative) replica 4.16 ± 0.14 2.04 ± 0.03 4.04 ± 0.02

A.F. Sartori, et al. Diamond & Related Materials 95 (2019) 20–27

26

nanoimprint lithography, Nanotechnology. 13 (2002) 592–596, https://doi.org/10.1088/0957-4484/13/5/309.

[27] Y. Hirai, S. Yoshida, N. Takagi, Y. Tanaka, H. Yabe, K. Sasaki, H. Sumitani,K. Yamamoto, High aspect pattern fabrication by nano imprint lithography usingfine diamond Mold, Jpn. J. Appl. Phys. 42 (2003) 3863–3866, https://doi.org/10.1143/JJAP.42.3863.

[28] M. Komori, H. Uchiyama, H. Takebe, T. Kusuura, K. Kobayashi, H. Kuwahara,T. Tsuchiya, Micro/nanoimprinting of glass under high temperature using a CVDdiamond mold, J. Micromechanics Microengineering. 18 (2008) 65013, https://doi.org/10.1088/0960-1317/18/6/065013.

[29] P. Fanzio, A. Cagliani, K.G. Peterffy, L. Sasso, High throughput soft embossingprocess for micro-patterning of PEDOT thin films, Microelectron. Eng. 176 (2017)15–21, https://doi.org/10.1016/j.mee.2017.01.011.

[30] P. Fanzio, C.-T. Chang, M. Skolimowski, S. Tanzi, L. Sasso, Fully-polymeric pHsensor realized by means of a single-step soft embossing technique, Sensors. 17(2017) 1169, https://doi.org/10.3390/s17051169.

[31] M. Tsigkourakos, T. Hantschel, S.D. Janssens, K. Haenen, W. Vandervorst, Spin-seeding approach for diamond growth on large area silicon-wafer substrates, Phys.

Status Solidi 209 (2012) 1659–1663, https://doi.org/10.1002/pssa.201200137.[32] C. Ramaswamy, Raman effect in diamond, Nature. 125 (1930) 704, https://doi.org/

10.1038/125704b0.[33] A. Merlen, J.G. Buijnsters, C. Pardanaud, A guide to and review of the use of

multiwavelength Raman spectroscopy for characterizing defective aromatic carbonsolids: from graphene to amorphous carbons, Coatings. 7 (2017) 153, https://doi.org/10.3390/coatings7100153.

[34] S. Prawer, R.J. Nemanich, Raman spectroscopy of diamond and doped diamond,Philos. Trans. R. Soc. London A Math. Phys. Eng. Sci. 362 (2004) 2537–2565,https://doi.org/10.1098/rsta.2004.1451.

[35] F. Andrea Carlo, J. Robertson, Raman spectroscopy of amorphous, nanostructured,diamond-like carbon, and nanodiamond, Philos. Trans. Math. Phys. Eng. Sci. 362(2004) 2477–2512, https://doi.org/10.1098/rsta.2004.1452.

[36] J.G. Buijnsters, L. Vázquez, Growth dynamics of nanocrystalline diamond thin filmsdeposited by hot filament chemical vapor deposition: influence of low sticking andrenucleation processes, J. Phys. Chem. C 115 (2011) 9681–9691, https://doi.org/10.1021/jp201919r.

A.F. Sartori, et al. Diamond & Related Materials 95 (2019) 20–27

27